Let us recall the distant past, what is the reason for us to abandon the small brick-like function machine, choose the smart phone? Let's take a look at the relevant content with the small series.
Is it because of the value? Interaction is fresh? I believe that most users, because the APP model brings too much practical value, people around are used, and they can't even follow up. Therefore, the greatness of Joe's gang is not only to subvert the shape of the mobile phone, but more importantly, to open the entrance to the ecology of the future mobile phone. To this day, its energy and imagination are far from exhausted.
By the time the mobile AI era kicks off, this logic seems to be universal. After all, the characteristics of the AI ​​algorithm are ridiculous, and everything is done.
Some people use AI for medical care, some use AI as customer service, and even some people use AI to predict the time of death. How to transplant these astounding ideas to mobile phones is obviously the real big business of this era.
But for the new artificial intelligence world, the introduction of developers and development ecology into their own platforms, thus forming the industrial barriers under the mobile phone AI proposition, is not easy in both technology and business. And this battle for the development of the ecology, in fact, has long been a hit in the giants.
For AI, millions of smart minds are worth
Looking at the many "AI mobile phones" that are released in this year, you will find many interesting phenomena. For example, Huawei's Mate10 first released the camera mode of scene recognition and data tagging. Last year, the flagship machines of all brands added similar functions.
Of course, this is not a bad function. Indeed, scene recognition + shooting can solve many problems and bring experience upgrades. But the problem is that this is too similar? How can AI with so many abilities become the same "student head" in the end?
In fact, the current high-achieving AI solution on mobile phones has dozens of capabilities in image recognition, environment understanding, image enhancement, NLP, and voice processing. At least improve the performance and experience of six major categories of live and short video, photography, social, shopping, AR, translation and other mainstream applications. The possibility of creating an unknown popular app is even more enticing. After all, mobile phone manufacturers themselves can never do much AI applications, and truly let the general public accept mobile AI, is a million smart minds, and even a new business model based on AI solutions.
But the fullness of the technical ideals that you want to shine into the application of the bones is always a little bit more distance.
For example, if there is a lack of AI special processing capability in the mobile phone, many AI tasks will be poorly put on the mobile phone, or simply cannot run. In the absence of platform and API openness, developers don't know how to port AI models to mobile phones. Farther away, developers can't rush into mobile AI to develop this new thing without being able to determine future benefits and business value.
Mobile phone AI itself calls for diversity of applications, and developers lack technical support and commercial guarantees. They dare not rush into mobile phone AI development, which has become the main contradiction of mobile phone AI topics. If it is not resolved, then we may face the embarrassing situation that only two or three new AI applications can be seen each year.
Of course, contradictions have always been opportunities. Especially for players who master the advantages of technology and the energy of the media, it is the best layout for the future when the industry is generally unable to complete one thing. Nowadays, the battle for mobile AI application development has actually quietly started among a few giants.
HiAI and TFlite: The ecological battle between the giants has already begun
For Chinese consumers, especially mobile phone enthusiasts, the most familiar mobile phone AI development platform today is definitely the HiAI mobile AI development architecture launched by Huawei after the Kirin 970.
Based on the terminal AI acceleration capability provided by Kirin 970, and Huawei and Glory have launched three products equipped with AI chips, the role of HiAI architecture is to open up the developer link and introduce ecological development capabilities.
So far, the HiAI architecture has been upgraded to version 2.0, which is compatible with almost all major deep learning development frameworks, and has introduced development tutorials and two generations of developer motherboards.
We are familiar with fast hands, vibrato, Mito, and many shopping and social applications, have already disclosed the cooperation with Huawei and HiAI architecture. For example, the fast hand will develop new live effects, gestures and limb recognition, scene recognition applications based on the HiAI architecture, and will also develop new compression models that can also use AI effects in weak network environments.
Similar to Huawei's idea of ​​implementing AI in the terminal, but it is very different, and it is probably the TensorFlow Lite that Google started at the end of last year. Unlike HiAI's Huawei-based AI chips and products, TFlite is essentially based on the deep learning development framework TensorFlow, but its purpose is to help developers develop and run learning models on local devices. This also leads to its characteristics biased towards the development side of the algorithm, while scorning applications and business.
However, according to related reports, some applications in Pixel 2 have been developed based on TFlite, and the Google mobile AI chip, which will appear in the future with great probability, will also be combined with TFlite. At present, many great gods on TensorFlow have made a lot of fun application models based on TFlite. These applications are far from Chinese users, but many ideas have reference value for domestic developers.
It is worth mentioning that although Apple has launched the AI ​​chip A11, it has not launched a platform-based product that supports the AI ​​ecological development. But this does not mean that Apple does not pay attention to the developer link. Last June, Apple opened up a machine learning capability called Core ML on the IOS development board. Currently, it has opened two API interfaces, Vision API and Natual Language API, which allow developers to develop machine vision and natural language processing based on this. Features.
At present, Apple is more inclined to small-scale low-level capability development, allowing developers to improve the application experience in the IOS environment, rather than making subversive development. According to the characteristics of Apple, he often prefers to hoard after the technology matures, with a superb engineering ability for one-time release.
In any case, the competition between the giants around the AI ​​development ecology has already set the direction. At present, the characteristics of their respective wars can be seen as the giants testing the water based on their own technical strength and strategic needs, which leads to the lack of uniform standards in the industry, but also brings opportunities for developers to grow wildly. In any case, the battle for the future, the core has long been determined.
What is mobile AI fighting for?
In the context of the beginning of new technologies to scouring demand and cognition, the platform and the developers in various fields want to integrate into the ecology, which is always determined by several factors.
The first priority is to reduce the barriers to entry.
This threshold includes many aspects, such as technical access thresholds; trial costs; learning finished products; migration costs; and compatibility costs. Developers can't spend too much time and money trying out unknown market propositions, and don't want to ruin themselves because of platform compatibility, framework migration, and so on. Developers who don't even know the algorithm, and want to enter this field to contribute wisdom and traffic. These are the responsibilities of the platform.
Second, good ecology is inseparable from a good empowerment program. Such as the distribution policy, market guidance links and so on. For developers, especially developers of high-competitive markets like China, business value is always a prerequisite. On the platform to provide good technical solutions, the platform must consider and effectively guide developers to obtain reasonable traffic and business returns.
Finally, in the case of weak market perceptions. A sensational “big move†case is perhaps more convincing than all technical parameters and market analysis. Once the platform has hatched a singularity application, the market will soon see commercial value based on AI chips and AI mobile development architecture. This confirms the reasonable way to enter the era of mobile phone AI. Logically we all know what the future will be, but in the real timeline, we must start a future by a certain case.
Overall, mobile AI ecological development is like a castle built in the future. We know that the flower is the truth of AI, and we know that the giants who hold the technological advantages have begun to break ground. Perhaps a little bit of time will be the final flavoring required for the entire AI mobile development.
For today's Chinese developers, it is clear that the HiAI architecture is a better choice. For developers with high-tech levels, it may be better to integrate different development solutions and form their own development system.
Perhaps an example can illustrate the value of AI development ecology for mobile phones: Today, a smart phone puts its own good, and the result is not equipped with WeChat, it is basically waste plastic.
By the same token, the value of an AI phone is to be able to carry an unknown AI application not far away. In this way, the strategic barrier of mobile phone AI is not at all a "AI" in marketing speech, but in the face of the development of ecological technology access threshold and ecological construction plan.
What we can see now, this is still a giant technology game. Of course, everything is still full of variables.
Basic Physics Experiment Instrument Series
Basic physics experiment instrument series, used in physics laboratories of colleges and universities.
Basic Physics Experiment Instrument,Light And Optical Instruments,Optical Viewing Instrument,Microscope Light Source Instrument
Yuheng Optics Co., Ltd.(Changchun) , https://www.yuhengcoder.com